Deep Learning for Automation of 3D Pore Analysis in Micro-CT Tomographs

This research proposal focuses on addressing the prevalent challenges associated with Proton Exchange Membrane Fuel Cells (PEMFCs), specifically targeting the reduction of greenhouse gas emissions. Despite significant advancements in performance and durability, particularly in automotive contexts, the high cost associated with essential materials for optimal functionality remains a formidable barrier. Notably, advancements in cathode gas diffusion layer (GDL) water management have enhanced performance metrics and subsequently mitigated costs. However, a detailed understanding of microscale pores and liquid water dynamics is crucial for further innovation. Current image processing methods for bubble analysis within GDLs are time-consuming, potentially biased, and limited in their capacity for complex analyses. This project proposes the integration of artificial intelligence (AI) to automate image analysis processes, aiming to increase precision, reduce bias, and expedite research timelines. By combining experimental insights with advanced image analysis techniques, this initiative strives to develop a comprehensive tool for characterizing and understanding the intricate phenomena occurring within PEMFCs. Collaborative efforts between theoretical and experimental research groups will facilitate the necessary data collection, development, and validation of this innovative tool, ultimately contributing to the advancement of PEMFC technology and its commercial application.

Faculty Supervisor:

Aimy Bazylak

Student:

Partner:

Rheinisch-Westfälische Technische Hochschule Aachen

Discipline:

Computer science

Sector:

Clean Technology; Environmental Science and Technology; Green/Alternative Energy

University:

University of Toronto

Program: